The effects of pre-task music on exercise performance and associated psycho-physiological responses: a systematic review with multilevel meta-analysis of controlled studies

This systematic review summarized the studies that examined the pre-task music effects on performance aspects and quantitatively analyzed their outcomes. A systematic search for controlled studies investigating the acute effects of pre-task music on physical performance, cognitive aspects and associated psycho-physiological responses was performed through Scopus, PubMed, Web of Science and Cochrane Library databases up to 17 May 2023, with thirty studies fulfilled the inclusion criteria. Data was analyzed using the robust multilevel meta-analysis model of standardized mean difference “SMD” with 95% confidence intervals (95%CI) and prediction intervals (PI) were reported. Pre-task music induced improvements of completion time (SMD = −0.24; 95% CI = −0.46 to −0.01; PI = −0.82 to 0.35; p = 0.04), relative mean power (RMP) (SMD = 0.38; 95% CI = 0.16 to 0.60; PI = −0.36 to 1.12; p = 0.003) and fatigue (SMD = −0.20; 95% CI = −0.32 to −0.09; PI = −0.36 to −0.05; p = 0.01), moderate effects on relative peak power (RPP) (SMD = 0.53; 95% CI = 0.21 to 0.85; PI = −0.42 to 1.48; p = 0.005), and high effect on feeling scale (FS) (SMD = 2.42; 95% CI = 0.52 to 4.31; PI = −11.43 to 16.26; p = 0.03). Greater benefits were recorded in jumping performance in males than females (p = 0.01), and for active than trained subjects for completion time (p = 0.02), RPP (p = 0.02) and RMP (p = 0.03). Larger benefits were obtained for FS post-warming up than after testing (p = 0.04). Self-selected music induced greater effects than pseudo- and pre-selected for performance decrement index (p = 0.05) and FS (p = 0.02). It could be concluded that pre-task music improved psychological responses and fatigue-related symptoms associated with exercise performance enhancement.


Introduction
The ergogenic effects of listening to music on performance have been well documented and were highly relevant to both practitioners and those operating in the sport domain (Terry et al., 2020).The effectiveness of listening to music has been mainly attributed to its benefits on the psychological state and fatigue-related symptoms associated with exercise performance and recovery status improvement (Karageorghis and Priest, 2012;Bigliassi, 2015).Regarding sport performance, music interventions may serve to enhance psycho-physiological responses and recovery both during and after exercise (Ka-Lok et al., 2020).However, practically, athletes may not readily be able to listen to music during competition and thus must listen to music before training/competition to modulate mental and perceptual issues (i.e., relief of stressors and competition-related anxiety) (Karageorghis et al., 2021).This pre-event process can help athletes to be prepared for the game through rehearsing association between imaging and relaxation skills (Gavanda et al., 2022).It has been demonstrated that pre-task music intervention improved power and strength (Smirmaul, 2017), distracted the intentional focus from discomfort in order to reduce the perceived exertion (RPE) and improved physical performance (Terry et al., 2020).However, inconsistent results about pre-task music effects have been reported (Smirmaul, 2017), with some studies showing improved power output (Jarraya et al., 2012;Chtourou et al., 2012aChtourou et al., , 2017)), running speed (Tounsi et al., 2019;Jebabli et al., 2022), kicking performances (Ouergui et al., 2023a,b) and reduced performance decrement index (PDI) (Puad et al., 2020;Jebabli et al., 2023;Ouergui et al., 2023a), while others (Bigliassi et al., 2012(Bigliassi et al., , 2015;;Loizou and Karageorghis, 2015;Gavanda et al., 2022) did not detect significant benefits even when music was self-selected (Fox et al., 2019).While there is no clear consensus about the underlying mechanisms related to performance changes after pre-task or warm-up music, collective evidence has suggested that motivation is a primary factor limiting the efficacy of pre-task and warm-up music (Smirmaul et al., 2014;Ballmann et al., 2019;Karow et al., 2020;Ballmann, 2021;Ballmann et al., 2021).This could implicate that various psychological factors largely affect subsequent performance and determine music intervention efficacy (Ballmann et al., 2021).To explain inconsistencies, psychophysiological factors could be differentially altered by intrinsic features of music (i.e., tempo, genre, intensity, exposure duration, synchronization), experimental procedures (i.e., selection process, intensity and duration of exercise) and the listener's characteristics (i.e., age, sex, training status) (Terry et al., 2020).
A recent meta-analysis about the effects of music in exercise and sports settings summarized data from pre, during and post-tasks interventions, resulting in high heterogeneous findings and limited their application (Terry et al., 2020).In fact, music interventions 'timing is a moderating factor which could limit the generalization of the aforementioned meta-analysis findings (Bigliassi et al., 2012;Jebabli et al., 2022), as sports governing bodies forbid electronic devices' use (e.g., personal music devices) during most game and competition scenarios (Chtourou et al., 2012b).Moreover, due to the increased body of evidence regarding the pre-task music effects on subsequent exercise responses with inconsistent findings being reported, conducting a meta-analytic review is highly pertinent in determining sound conclusions.Therefore, this systematic review and meta-analytic study was designed to summarize reports about pre-task music effects on exercise performance and related responses, metaanalyze their data taking into consideration the different moderators that could explain the variability across studies.Moreover, the study tends to highlight the limits of the existing literature and draw some practical recommendations that could improve the effectiveness of using pre-task music stimulation.

Materials and methods
This systematic review was conducted following the 2020 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (Page et al., 2021) and the PRISMA implemented in Exercise, Rehabilitation, Sport medicine and SporTs science (PERSiST) guidelines (Ardern et al., 2022).The protocol for this systematic review was published in the Open Science platform (OSF) on July 7, 2023.Registration DOI:10.17605/OSF.IO/XHBN4.

Search strategy
PubMed/MEDLINE, Scopus, Web of Science, Sport Discus, and Cochrane Library were searched without date limits or filters.The search was performed up to 17 May 2023 connecting the terms using the Boolean operators ' AND' and 'OR' with medical subject heading (MeSH) terms appropriately used.The search strategy on each database was presented in the Supplementary Table S1.The literature search was expanded further by hand searching review articles on the topic and by reference lists' searching from the articles retrieved for relevant references.

Inclusion and exclusion criteria
The PICO (i.e., population, intervention, comparator, and outcome) model was used in studies inclusion: • Population: Healthy subjects.
• Intervention: Acute effects of isolated music intervention during warm-up or before exercise.• Comparator: No music stimulus (control).
These eligibility criteria resulted in excluding studies that were not associated with the established PICO criteria.Additionally, studies were excluded if music intervention effect could not be isolated from other stimulation (e.g., accompanying video footage, or brainwave synchronizer), if a case study or a longitudinal study design have been used, if a clinical or special population has been studied, or if a subjective measure of performance was used (e.g., interview).Moreover, books, citations, trial registry records, conference proceedings, systematic reviews, and narrative reviews were excluded.

Study selection
The retrieved articles were first assessed for duplication using the software "Endnote 20" (Camelot UK Bidco Limited-Clarivate, United Kingdom) before being considered for inclusion.After duplicate removal, relevant articles' titles review was conducted before examining articles' abstracts and then fully-published peer-reviewed articles.Two reviewers conducted the process independently and consensus was used to resolve disagreements between them.The reasons for the exclusion of full-text articles were recorded.

Data extraction
For studies meeting inclusion criteria, data were summarized in a Microsoft Excel spreadsheet.A piloted data extraction form with the following items: author(s), year of publication, sample size and sex, sample age, exercise performed, main outcomes, exposure duration, music selection, music tempo, loudness and the main results were used.For the main results, all data about the effect of pre-task music on physical performance, perceived exertion, fatigue symptoms, affective and physiological responses before, during or/and following exercise were extracted from the papers.All eligible outcomes including repeated measurements at different time points were considered as one unit of evidence and coded accordingly.If data were not reported in a way that was conducive to extraction for our analysis, we request appropriate data from respective authors [i.e., mean (standard deviation), raw data].If authors did not respond our request, we extracted relevant data in studies that only reported graphical information (Drevon et al., 2017) using WebPlot Digitizer (v4.3, Ankit Rohatgi).The data extraction process was double-checked by a second reviewer to avoid any selection bias and data extraction flaws.

Moderator variables coding
A wide variety of moderating factors were tested and coded as presented in Table 1.The included studies were coded for music characteristics (i.e., selection, tempo, exposure duration), study characteristics (i.e., time of experimentation, timing of variable measurement, test duration), and participant characteristics (i.e., sex, age, and training status).

Music characteristics
Music selection Self-selected: music stimuli selected by participants Pre-selected: music stimuli selected by the experimenter(s) Pseudo: music stimuli selected by participants from an experimenter-defined list or the opposite

Music tempo
Fast: refers to music with a tempo >120 bpm Slow-to-medium tempo: refers to music with a tempo <120 bpm Exposure duration Kept as a continuous moderator

Participants characteristics
Training status Trained: refers to participants who engaged in regular physical activity (3 times/week)

Assessment of risk of bias
The methodological quality of the included studies was assessed using the Physiotherapy Evidence Database (PEDro) scale (Maher et al., 2003;de Morton, 2009).Given that blinding in the traditional experimental sense is not possible with a music treatment (Terry et al., 2020), items 5-7 (which are specific to blinding)were removed from the scale (Garbisu-Hualde and Santos-Concejero, 2021).After these items' removal, the maximum result on the modified PEDro scale was 7 (i.e., the first item is not included in the final score)and the lowest was 0 (Garbisu-Hualde and Santos-Concejero, 2021).The quality assessment was interpreted as follows: poor quality (≤3 points), moderate quality (4-5 points), and high quality (6-7 points).Two reviewers performed the methodological quality assessment independently.

Effect sizes calculation
The standardized mean difference "SMD" was calculated from primary studies by subtracting the means and SDs from music and no music conditions, the sample size and the raw correlation coefficient.For within-participant effects, pre-post correlations for measures have not been reported, therefore, we adopted a range of values for correlation coefficients (r = 0.5, 0.7, and 0.9) and examined the sensitivity of the results to each of these values.Since the overall results were relatively insensitive to this range, we reported the results for r = 0.7 here and included the results for the other assumed correlation coefficients in the Supplementary File.

Overall effects synthesis
The effect sizes' independence is crucial to avoid an information overlap and to provide unbiased effect size's estimations (Cheung, 2014;Gucciardi et al., 2021).Multilevel meta-analysis has been reported as a valid and efficient way to handle dependent effect sizes within a study (Assink and Wibbelink, 2016;Moeyaert et al., 2017).Most studies in the present systematic review are multi-arms.Therefore, to deal with the statistical dependency caused by multiple observations inclusion from the same study, a multi-level, randomeffects model for meta-analysis was performed using the "metaphor" package (rma.mvfunction) (Viechtbauer, 2010) and complemented by a cluster-robust inference method (Pustejovsky and Tipton, 2022) to generate an overall mean effect size (ES) and 95% confidence interval (95%CI).Furthermore, 95% prediction interval (PI) was calculated since it serves to determine the expected range of ES in future primary studies (IntHout et al., 2016).A multilevel metaanalysis was conducted for every outcome separately and model parameters were estimated by the restricted maximum likelihood estimation method.
Exploratory subgroup comparisons of moderator variables were performed, including music selection, tempo, sex, training status, timing of variable measurement, test duration, and time of day.Metaregressions were calculated for mean age, and exposure duration to the music.Moderator analysis was performed when there were sufficient observations per group (i.e., at least 3 observations) (Spruit et al., 2016) and robust estimates were produced.

Publication bias
The risk of small study bias was visualized through contourenhanced funnel plots.Moreover, multilevel extension of the Egger's test was used to detect publication bias.The variance of effects was converted into standard error and the moderation model was performed (Fernández-Castilla et al., 2021).The robust estimates results of the Egger's test were produced for all analyses.

Outlier detection
For all meta-analytic models, outlier and influential case diagnostics were performed by calculating Cook's distance and standardized residuals, respectively (Cook, 1977;Viechtbauer and Cheung, 2010).Cases were considered outliers when their Cook's distance's values greater three times than their respective mean, and with a standardized residual value greater than 3, in absolute values (Gucciardi et al., 2021).In the case of existing outliers, the overall effects were recalculated to assess the robustness of the fitted model.

Search results
A summary of the search process is shown in Figure 1.Search strategies identified 482 citations related to music in physical activity and exercise.After 184 duplicates' removal, title and abstract of 298 citations were screened and 38 studies were targeted for detailed review (i.e., 260 excluded).In total, 11 studies which did not meet all inclusion criteria were excluded after full-text screening.The excluded studies and exclusion reasons were listed in the Supplementary Table S2.
After an additional search on Google scholar and in the reference lists of the relevant studies, three published articles were added as relevant papers, resulting in a net number of 30 studies from the overall search process.

Methodological quality
By removing blinding criteria, included studies presented a moderate to high methodological qualities (5-7 points) as presented in Table 3.The primary issue with the Pedro items is the randomization procedure, which was not covered in depth in the included researches.
Funnel plot of per-study standard error by SMD (Supplementary Figure S1) indicated significant asymmetry.The robust multilevel model of Egger's test showed potential risk of publication bias (F 1, 1.39 = 228.27,p = 0.02).
Funnel plot of per-study standard error by SMD (Supplementary Figure S2) indicated significant asymmetry.The robust multilevel model of Egger's test showed a potential risk of publication bias (F 1,2.02 = 45.18;p = 0.02).
Funnel plot of per-study standard error by SMD (Supplementary Figure S3) indicated significant asymmetry.The robust multilevel model of Egger's test showed a potential risk of publication bias (F 1,2.18 = 68.73;p = 0.01).
Funnel plot of per-study standard error by SMD (Supplementary Figure S4) indicated significant asymmetry.Furthermore, the robust multilevel model of Egger's test showed a potential risk of publication bias (F 1,2.25 = 46.02;p = 0.02).
Funnel plot of per-study standard error by SMD (Supplementary Figure S5) indicated significant asymmetry.The robust multilevel model of Egger's test showed potential risk of publication bias (F 1, 1.16 = 1380.76;p = 0.01).
Funnel plot of per-study standard error by SMD (Supplementary Figure S6) indicated no significant asymmetry.The robust multilevel model of Egger's test showed no potential risk of publication bias (F 1,4.22 = 3.45; p = 0.13).
Funnel plots of per-study standard error by SMD (Supplementary Figure S8) indicated no significant asymmetry.The robust multilevel model of Egger's test showed no potential risk of publication bias (F 1, 3 = 5.16; p = 0.11).
Outlier and influential case diagnostics were performed and showed no outlier residuals or Cook's outliers.

Discussion
This is a meta-analytic review about pre-task music effects on physical performances and the associated psycho-physiological aspects.The meta-analysis showed that pre-task music was associated with significant improvement in completion time, relative mean and peak powers, affective valence, and fatigue symptoms.Moderators' analysis revealed that sex was a significant factor that affected jump height, with greater performance in males than females.Training status modulated the physical performance with completion time, RPP and RMP were greater in active than trained subjects.In addition, variable measurement timing moderated pre-task music effect on FS with larger benefits post-warming up than after testing.Furthermore, music selection process was a significant moderator of PDI and FS with greater effects from SSM than PSM and pseudo selected music.

Main effects
The present study showed that pre-task music improved different aspects of both physical and psychological aspects of responses to exercise.The findings of the present systematic review with metaanalysis supported previous meta-analyses which suggested that music was associated with significant beneficial effects for physical performance (Terry et al., 2020) and that when outcomes were expressed relatively to body mass (i.e., RPP and RMP), listening to music induced significant moderate-sized effects (Castaneda-Babarro et al., 2020).According to Karageorghis and Terry (2011), listening to the appropriate music may boost exercise performance up to 15%, which is comparable to commonly used supplemental ergogenic aids (Karageorghis and Terry, 2011).It has been shown that music acts as an ergogenic aid which improves exercise performance by attenuating exhaustion or increasing work capacity (Karageorghis and Priest, 2012).A decrease in power output over the course of total effort time during a task reflects a fatigue index (Castaneda-Babarro et al., 2020).In the present meta-analysis, the effect of pre-task music on PDI was moderate but non-significant.The observed effect could be practically important in sport settings since seemingly trivial performance improvements for an athlete may be decisive for the event's outcome (Hopkins et al., 1999).The moderate PDI decrease might be explained by the energy economy improvement (Atkinson et al., 2004).However, it is crucial to keep in mind that PDI largely depends on the muscle fibers' type of participants, as a larger percentage of muscle fibers enable higher peak power but a higher fatigue rate (Castaneda-Babarro et al., 2020).Since music stimulates the motor cortex and other regions of the brain involved in the control of movement, its effects on power output may be related to changes in neuromuscular activity (Gordon et al., 2018).This could be evident since pre-task music has showed a likely ergogenic effect on short and predominantly anaerobic tasks (Smirmaul, 2017).Physiologically, warm-up music has been shown to increase catecholamine release which may partially explain favor to anaerobic tasks (Yamamoto et al., 2003).This could be supported by the present meta-analysis results where pre-task music presented high practical effect on jumping performance even it was not statistically significant.Possibly, due to the short duration and volume of exercises performed after listening to music that fatigue was lower to alter performance when music effects dominate (Bigliassi, 2015).Alternatively, being efficient in explosive movements is based on an optimal psychological state which may be enhanced by listening to music even when it is performed prior to performance (Ballmann et al., 2021).
Regarding the importance of a positive psychological state during exercise, music was considered as an efficient strategy to increase sport activities commitment, even performed at high-intensities (Clark et al., 2015;Hutchinson et al., 2018) and may positively affect performance (Castaneda-Babarro et al., 2020).The present metaanalysis highlights fatigue symptom decreases and that affective valence scores were directed toward the positive end of the scale when pre-task music was used.The positive effect of music toward affective valence supported previous findings reporting moderate effect of music on FS (Terry et al., 2020).The use of simulative or motivational music implicates the brain stem reflex, stimulating the central nervous system in a manner reflecting physiological arousal increases which is associated with high-intensity activity (Karageorghis and Jones, 2014).Furthermore, when listening to music, the phenomenon of emotional contagion may occur in a manner that involves the exerciser or athlete to recognize the music's emotional elements (Castaneda-Babarro et al., 2020).Bishop et al. (2013) indicated that listening to motivational music stimulates the primary auditory membrane parts and cerebellum responsible for processing emotions, managing movement or movement patterns control.Their mutual stimulation with music can be one of the main reasons for their effectiveness in listeners' performance (Bishop et al., 2013).
Music has been shown likely to enhance motivation and arousal for sports (Karageorghis and Terry, 1997), which are the underlying psychological factors to maximize physical effort and performance (Anshel, 2019).In the present systematic review, evidence has consistently supported the idea that listening to music improves motivation to exercise (Smirmaul et al., 2014;Karow et al., 2020;Ballmann et al., 2021).Smirmaul et al. (2014) showed that trained swimmers who listened to music during warm-up felt more motivated prior to exercise and subsequently swam faster in the 200 m trial.From the other side, Karow et al. (2020) reported higher power output and speed in a rowing exercise related to higher motivation after listening to preferred music during warm-up.Such motivational effects of pre-task or warm-up music may not only be limited to preexercise, but this motivation may last throughout the exercise bout (Ballmann et al., 2021).Therefore, motivation enhancement reported by these previous studies likely led to increased effort and improved velocity and power output (Ballmann, 2021).As well, the ability to increase arousal and calm down was clearly associated with pre-task music (Bigliassi et al., 2015).Interestingly, previous meta-analysis showed the effectiveness of music to lower both physiological stressrelated arousal (e.g., blood pressure, heart rate, and hormone levels) and psychological stress-related experiences (e.g., state anxiety, restlessness or nervousness) in various populations and settings (de Witte et al., 2020).Moreover, another meta-analysis showed that listening to music induced an overall significant large effect on alleviating anxiety (Harney et al., 2022).Such results had received consensus from pre-task music investigations where music reduced anxiety (Aloui et al., 2015), negative mood states (Khemila et al., 2021;Gavanda et al., 2022) and increased vigor (Chtourou et al., 2017).From a physiological point of view, pre-task music was associated with reduced post-exercise cortisol (Khemila et al., 2021) and maintained heart rate values (Jarraya et al., 2012;Arazi et al., 2015;Bigliassi et al., 2015;Jebabli et al., 2022Jebabli et al., , 2023) ) despite performance improvements.Decreased cortisol concentration may serve as an underlying mechanism for reducing fatigue (Ghaderi et al., 2015).Moreover, improved parasympathetic recovery (Heart rate variability analysis) was found to be the cause of music calming effect (Bigliassi et al., 2015).
Affective responses to exercise are jointly influenced by cognitive factors, such as physical self-efficacy and interoceptive (e.g., muscular or respiratory) cues that reach the affective centers of the brain via sub-cortical pathways (Ekkekakis, 2003).Using stimulating music, 10.3389/fpsyg.2023.1293783Frontiers in Psychology 13 frontiersin.orgevidence shows improved cognitive performances during simple and complex tasks (Bishop et al., 2013;Jarraya and Jarraya, 2019;Khemila et al., 2021).The improvement of reaction time, attention allocation, and visual perception were the cognitive skills connected with pre-task music effects.From a neurological approach, music yielded persistent activation in the basal ganglia, which is implicated in effective decision-making and preparation for action in sport (Yarrow et al., 2009).In addition, the use of fast and loud music induces a brain state whereby an individual vigorously monitors his/her environment, and swiftly identifies relevant targets (Bishop et al., 2013).
Regarding RPE, the significant influence of music on perceived exertion was explained primarily by its ability to distract exercisers from unpleasant, fatigue-related sensations (Terry et al., 2020).However, using pre-task music stimulations, the current meta-analysis revealed a trivial non-significant effect on RPE.Although the benefits of music on RPE have been observed even during high-intensity activity (Bigliassi et al., 2012;Bentouati et al., 2022), its distraction effect can be denied by powerful interceptive signals of physical discomfort associated with the activity and the benefit to RPE may be lost (Stork et al., 2015).It is possible that, to maximize these benefits of shifting an external focus, music may need to be played simultaneously with, rather than before, the exercise task (Ballmann et al., 2021).This effect of the listening time could explain the significant effect of music on RPE when including data from during exercise (Terry et al., 2020).The effect of pre-task music has been reported to decrease over time due to fatigue's increase (Aloui et al., 2015;Bigliassi, 2015).In fact, fatigue's increase is generally stronger than the music effects, which means that it is only "a matter of time" until fatigue-related symptoms overcome the ergogenic effects of music (Rejeski, 1985;Bigliassi, 2015).Therefore, the luck of music analgesic effects could be supported by the intensity dependent theory (Bigliassi et al., 2012).Since effort intensity is often expressed as a percentage of maximal oxygen uptake, it seems that music reduces perceived exhaustion only up to 75% of maximal oxygen consumption, and these effects disappear quickly when intensity is above (Hutchinson and Karageorghis, 2013).Therefore, RPE values are likely maintained because self-pacing enables people to exercise at a determined RPE level throughout exercise (Fox et al., 2019).

Moderating variables
Situational and inter-individual variables are the mediators affecting the power of communication between a piece of music and the individual's response (Hutchinson and Karageorghis, 2013).However, Terry et al. (2020) suggested that music is likely to induce positive responses during exercise on a fairly consistent basis regardless of personal, situational and musical characteristics.In the present meta-analysis, pre-task music effect on affective valence seems to be modulated by the selection process and the timing of variable measurement.The promotion of self-determined forms of behavioral regulation are likely to encourage the maintenance of physical activity behaviors (Ryan and Deci, 2017;Terry et al., 2020).Specifically, the purpose of selecting music is, generally speaking, to ensure that the individual objective is optimized (Eliakim et al., 2012).As a result, the music type that an individual can choose is of importance to determine the psycho-physiological impact of music on physical performance capacities (Işık et al., 2015).Jiang et al. (2016) showed that listening to self-selected music elicited stronger and more positive emotional responses regardless of the song's valence (i.e., positive or negative) and arousal (i.e., high or low).This might be attributed to the fact that self-selected music was more preferred and familiar (Jiang et al., 2016).In fact, neural activation changes elicited by listening to music have been shown to be heavily dependent on the individual's musical preference where greater emotional connection to preferred music may alter responses (Holler et al., 2012).In other words, listeners may respond differently to music melody and harmony based on their cultural background and familiarity level (Akhshabi and Rahimi, 2021).The higher performance under the preferred condition was also linked to expectations influence on the recorded effects (Ouergui et al., 2023b).
Individual factors such as sex (Nater et al., 2006;Karageorghis et al., 2018) and training status (Carlier et al., 2017) have been suggested to modulate the reaction to music during exercise.The present study revealed that music resulted in greater effects within males than females during explosive tasks.More potent effects of music in males could be related to their ability to develop higher muscular power while maintaining lower fatigue rate than females (Ouergui et al., 2023b).Regarding psychological responses, females may exhibit higher emotional sensitivity to musical stimuli compared to males (Nater et al., 2006).For instance, females fixate more attention to music's rhythmic quality and movements than males, and males to the cultural bond of music more than their female counterparts (Karageorghis et al., 1999).Consequently, since the majority of examined studies used pre-selected music, male responses were positively altered maybe due to their higher objectivity (Nater et al., 2006).Considering training status, the influence of music on performance was reported to decrease significantly with increased fitness level (Eliakim et al., 2007).There is evidence that highly trained exercisers or elite athletes tend to associate rather than dissociate (Gabana et al., 2014) and then their RPE scores decrease due to music, even at high work intensities (Terry et al., 2020).In the present metaanalysis, the effects of pre-task music on physical performance outcomes (i.e., completion time, RPP and RMP) were greater on active than trained, whereas its effect on affective valence was greater in trained than active subjects.This result contradicts what was reported by Smirmaul (2017) where pre-task music effects were consistent regardless the fitness levels of individuals (i.e., varied from physically active students to national-level athletes).Our results could be attributed to the fact that trained subjects reached their top physical fitness level and then cannot benefit from music ergogenic effect (Berthelot et al., 2015).Thus, our findings could suggest that in highly trained subjects, pre-task music benefits on physical performances are highly mediated by its ergogenic potential on affective states (Carlier et al., 2017;Ballmann, 2021).This might be supported by the fact that music brought greater pleasure to the high than to the low tolerant participants (Carlier et al., 2017).
Regarding music intrinsic features, Terry et al. (2020) reported that fast-tempo music was associated with greater benefits than slowtempo music.This is reasonable, since people express a preference for keeping the tempo relatively high during intense exercise (Thaut, 2013).Given the high-energy/activation state typically required for optimal performance in exercise or sport, the stronger effect of fasttempo music reflects what we know about physiological arousal and musical aesthetics (Karageorghis, 2020).Yamamoto et al. (2003) showed that slow-rhythm music decreased excitement, while fast-rhythm music increased it.This rise in physiological excitement might be related to upbeat music ability to boost the heartbeat immediately before activity (Eliakim et al., 2007;Puad et al., 2020).Fast-rhythm music automatically stimulates the listener by activating the central nervous system, regardless of how the music is evaluated (Van Dyck, 2019).However, the present meta-analyses results did not support the moderating effects of music tempo as reported by Terry et al. (2020).It is important to note that when the power of physiological feedback signals overshadows the attention processes, the music ability to maintain physiological effects in strenuous physical activity is greatly altered (Ekkekakis, 2003).Since music is played before task execution, its tempo moderating effects may be dependent on listener' affective memory and dissociation ability (Bishop et al., 2013;Stork et al., 2015).

Strengths and limitations
This is the first systematic review with meta-analysis addressing the acute effects of pre-task music on physical performance, cognitive aspects, and psycho-physiological responses associated to exercise and sport fields.The present investigation includes a comprehensive exposure of the available literature and a careful assessment of its importance as well as the presentation of the results from trials with moderate to high methodological quality.In addition, meta-analysis was conducted using a robust estimate model with the inclusion of all the available observations, which is the recommended approach to deal with dependency and show unbiased results (Assink and Wibbelink, 2016;Moeyaert et al., 2017).However, some limitations should be acknowledged in the present systematic review.In fact, due to the lack of sufficient information about music' characteristics (e.g., intensity of music, type of music), it is impossible to make perfect subgroups comparisons.Moreover, a new aspect observed in this study is that these studies (Bishop et al., 2013;Smirmaul et al., 2014;Jarraya and Jarraya, 2019;Ballmann et al., 2021;Gavanda et al., 2022) used pre-task music at rest rather than during the warm-up, raising a topic for further discussion.Another limitation of the included studies was the lack of controlling/reporting methodological procedures in sufficient details.For instance, information regarding music's characteristics such as intensity and the time between music exposure and task initiation was not sufficiently acknowledged.Regarding the mechanisms behind pre-task music effects, none of the included studies has directly measured dissociation, leaving its contribution to the music modulating effect on performance unclear at this time.Furthermore, despite motivation was the most common mechanism to explain warm-up music ergogenic potential, only three studies (Smirmaul et al., 2014;Karow et al., 2020;Ballmann et al., 2021) measured motivation level while the other ones take it as evident.

Future research perspectives
Pre-task music application remains restricted to exercise domain and less investigated in sport events (i.e., luck of investigations with competitions).This is why it would be appropriate to conduct further studies about the ergogenic effect of music on sport events.Moreover, even specificity is crucial training principle, it is clear that most of studies about pre-task music effects used generic rather than specific tests, which may limit generalization of results to specific conditions.Moreover, pre-task music exposure at rest vs. during the warm-up could generate different responses.In addition, determining the optimal time between pre-task music exposure and beginning of the task could be an interesting factor in music topics.Furthermore, as the objective in sports settings is to improve performances as much as possible, combining pre-task music with other ergogenic aid would be an attractive strategy.

Conclusion
As a psycho-active stimulus, warm-up music could be an effective strategy to enhance several aspects of performance (predominantly anaerobic tasks were the most affected), associated psychological responses and delay the fatigue-related symptoms.Such benefits of warm-up music could present implications on competition and training settings.Specifically, as a way to deal with psychological stress anticipated before competition, warm-up music could be highly attractive for elite athletes where minor differences determine the winner.Moreover, starting training sessions with high level of motivation and power help to achieve training objectives with lower cost, especially, for less trained subjects.However, pre-task music is a sensitive stimulus, which could be influenced by intrinsic and situational factors.Coaches seeking to improve their athletes' performances and prepare them to cope with training and competitions demands using pre-task music are recommended to give them the opportunity to select their preferred music since subjects' characteristics (e.g., sex, training status) seems to be the most moderating factors with varied preference level.In addition, since warm-up music effects depends on listeners' sex, findings from one sample sex should not be generalized to the other one and taken with caution.Moreover, users are recommended to improve the quality of their music selection process by considering the music's motivational properties and working with a guiding rationale.Even under preferred conditions, controlling the intrinsic components (e.g., volume, genre, melody) of the selected music could help to explain, in part, the varied responses across listeners.Furthermore, explaining warm-up music benefits should be articulated on measured responses (e.g., physiological, neurological, and psychological) that should be taken into consideration to explain the inter-individual variability and the existence of responders and non-responders to the stimulus.
FIGURE 1Flow diagram of the search process.

TABLE 1
Moderating factors and their codes.

TABLE 2
Studies characteristics.

TABLE 3
Pedro scale items (I) and studies' scores.